Learning Geocentric Object Pose in Oblique Monocular Images
This addresses a problem in Earth observation and satellite image analysis where traditional depth estimation methods fail, enabling better object detection and image alignment for remote sensing applications.
The paper tackles the challenge of estimating object geocentric pose (height above ground and orientation) in oblique monocular satellite images where depth cannot be reliably estimated, by developing a deep network trained on airborne lidar data. The approach dramatically improves localization accuracy and enables accurate alignment of images from different oblique viewpoints, as demonstrated on two large-scale public datasets for semantic segmentation.
An object's geocentric pose, defined as the height above ground and orientation with respect to gravity, is a powerful representation of real-world structure for object detection, segmentation, and localization tasks using RGBD images. For close-range vision tasks, height and orientation have been derived directly from stereo-computed depth and more recently from monocular depth predicted by deep networks. For long-range vision tasks such as Earth observation, depth cannot be reliably estimated with monocular images. Inspired by recent work in monocular height above ground prediction and optical flow prediction from static images, we develop an encoding of geocentric pose to address this challenge and train a deep network to compute the representation densely, supervised by publicly available airborne lidar. We exploit these attributes to rectify oblique images and remove observed object parallax to dramatically improve the accuracy of localization and to enable accurate alignment of multiple images taken from very different oblique viewpoints. We demonstrate the value of our approach by extending two large-scale public datasets for semantic segmentation in oblique satellite images. All of our data and code are publicly available.